COMMENT AND DEBATE

COMMENT AND DEBATE

What can the life course approach contribute to an of risk?

David Blane ESRC International Centre for Life Course Studies in Society and University College London, UK [email protected] Bola Akinwale ESRC International Centre for Life Course Studies in Society and Health and Imperial College London, UK Rebecca Landy ESRC International Centre for Life Course Studies in Society and Health and Queen Mary College London, UK Katey Matthews ESRC International Centre for Life Course Studies in Society and Health and University of Manchester, UK Morten Wahrendorf ESRC International Centre for Life Course Studies in Society and Health and University of Duesseldorf, Germany Hans-Werner Wahl Heidelberg University, Germany Mark D. Hayward University of Texas at Austin, USA Aart C. Liefbroer Netherlands Interdisciplinary Demographic Institute, University Medical Center Groningen and Vrije Universiteit Amsterdam, Netherlands Gita D. Mishra University of Queensland, Australia Isabel Ferreira University of Queensland, Australia Ilona Koupil Stockholm University Karolinska Institute, Sweden

http://dx.doi.org/10.14301/llcs.v7i2.343

Abstract means living longer than predicted. Attempts to understand longevity risk to date have concentrated on single , usually coronary heart , and sought explanations in terms of change and medical innovation. In an opening paper, David Blane and colleagues point to evidence that suggests changes in positive health also should be considered; and that a life course approach can do so in a way that is socially and biologically plausible. Applying this approach to UK citizens currently aged 85 years suggests that life course research should give priority to trajectories across the whole life course and to the social and material contexts through which each cohort has passed. Testing these ideas will require inter-disciplinary and international comparative research. The opening paper is followed by commentaries from Hans-Werner Wahl, Mark Hayward, Aart Liefbroer and Gita Mishra. Finally Blane and colleagues respond to the points raised by the commentators.

Keywords Longevity risk, positive health, social and biological plausibility, life course trajectories, social context

165

COMMENT AND DEBATE

Introduction or the introduction of new medical treatments. The increase in at middle age was Positive health, in the sense of growth and one of the defining characteristics of the late development, functional capacity, vitality and twentieth century, with all-cause mortality around resilience, was largely ignored. The present paper the state pension age in England and Wales falling by attempts to rectify this neglect by arguing that some two thirds during 1971-2001 (Akinwale et al., positive health should be considered alongside risk 2011), and similar improvements occurring in many factor change and medical innovation as driving the other countries (World Health Organisation, 2008). increase in life expectancy at middle age; and that The actuarial profession led research into the causes the life course approach can contribute to the of this change. Their thinking includes the idea of the understanding of longevity risk by bringing both golden cohort, born in the 1930s, who were the first biological and social plausibility (Blane, Kelly-Irving, in the UK to experience the fall in middle aged d’Errico, Bartley & Montgomery, 2013) to the idea of mortality (Willets, 2004), and the concept of positive health. longevity risk to describe the phenomenon of living Two pieces of evidence support the idea of positive longer than predicted (Willets et al., 2004). health. Table 1 shows for England and Wales, the The rate of improvement in middle-aged mortality disease-specific mortality rates for the most was not anticipated, with serious consequences for prevalent causes of death, which together account the financing of pension arrangements. In the UK, for some two-thirds of all deaths. Between 1971 and the market for annuities collapsed, employers closed 2001 their mortality rates tended to fall by defined benefit pension schemes for much of the proportionately similar amounts. The tendency was private sector and the age at which the state pension stronger for men than for women. Male mortality is paid was increased first by five years for women, to rates for all of these prevalent causes of death be followed by a further increase of three years for reduced by around 50-75%. For women, in contrast, all citizens. the fall in mortality rates was in the 50-75% range for Most of the actuarial and epidemiological research only five of the prevalent causes of death, while into the causes of this increasing longevity is disease- another reduced by around one-third and two specific, usually coronary heart disease (Unal, increased (it is tempting to attribute the rise in Critchley & Capewell, 2005; Bajekal et al., 2013) or women’s COPD and lung mortality to their cerebrovascular disease (Raine et al., 2009); and tobacco smoking, although the fall in their CHD assumed that the main determinants of the falling mortality, which is also tobacco-related, gives pause mortality rates would be either change in risk factors for ).

Table 1. Age-standardised mortality rates per million population; England & Wales, 1971-2010; selected causes, by prevalence.

Males Females 1971 2010 Percentage 1971 2010 Percentage change change 1971-2010 1971-2010 Ischaemic heart disease 3,801 1,083 -71.5 1,668 478 -71.3 Cerebrovascular disease 1,541 422 -72.6 1,352 396 -70.7 Chronic obstructive 944 314 -66.7 193 212 +9.8 pulmonary disease Pneumonia 920 260 -71.7 624 212 -66.0 Carcinoma of lung 1,066 465 -56.4 183 299 +63.4 Accidents and violence 333 170 -48.9 166 53 -68.1 Carcinoma of breast 4 2 -50.0 379 245 -35.4 Carcinoma of stomach 317 72 -77.3 149 33 -77.9 Source: Office for National Statistics (2011).

166 COMMENT AND DEBATE

Something similar happened in the late Figure 1 shows the change in the death rates of nineteenth and early twentieth centuries when people aged 65 years in England and Wales from mortality due to most of the prevalent infectious the 1840s to 2010. The mortality rates of women diseases of childhood fell at around the same time and men changed little (if anything they increased) (McKeown & Lowe, 1966; McKeown, 1979), during the nineteenth century from 1841-1849 to although in that instance the situation was simpler 1891-1899. After 1900 women’s at conceptually because infectious disease age 65 years fell in an approximately linear fashion already contained the idea of host across the whole of the twentieth century. Men’s resistance. Positive health can be seen as the mortality in contrast, after an initial fall, plateaued chronic equivalent of host from 1931-39 to 1961-69 followed by a precipitous resistance. fall to near-equality with the women’s rate by 2011.

Figure . Death rates of men and women at age years per live population of that age in England and Wales to

Source: Human Mortality Database.

The century-long, near linear fall in women’s requires explanation (the plateau coincided with mortality may have received too little attention. It the introduction of antibiotics and widened access suggests a cumulative process driven by rising living to, and sizeable investment in, medical care), with standards, with little sign of any large impact from the latter a process of catching up for lost time. the introduction of antibiotics, the start of the If it is accepted that risk factor change and National Health Service or the changing prevalence medical innovation have difficulty, on their own, in of tobacco smoking; also, it is easier to predict accounting for the twentieth century’s near linear future developments, including the necessary level fall in the mortality of 65 year old women and the of pension contributions, from linear change. proportionately similar fall in mortality from most Prediction would have been more difficult from the of the prevalent causes of death, then it may be change in men’s mortality at age 65 years: first the worth considering positive health as a third factor in three-decade plateau from the 1930s to the 1960s the explanation of longevity risk. The next section of suggested stability rather than change, then the the present paper attempts to do so by examining subsequent precipitous decline suggested rapid the biological and social life course of the golden change. On one reading, it is the former that cohort, operationalised as those currently aged 85

167 COMMENT AND DEBATE years. In both cases, biological and social, the Biological Life Course various life course stages of these 85-year-old Figure 2 presents a model of the biological life persons will be located within their UK social history course. The model was described by Strachan and context. Sheikh (2004) in relation to lung function; here it is generalised to positive health.

Figure 2.

Source: based on Strachan and Sheikh (2004), own visualization.

First there is a phase of growth and and residential crowding, is associated with late development, during which an initial fertilised cell adolescent sub-clinical abnormality (erythrocyte divides and replicates to an estimated 37.2 trillion sedimentation rate, proteinuria, blood pressure, cells by around age 20 years (Bianconi et al., 2003). body mass index, vulnerability to stress) which, in Functional capacity subsequently is lost: slowly at turn, is associated with premature death in middle first, when the loss is more than compensated by age (Sundin, Udumyan, Sjostrom & Montgomery, experience, anticipation and conditioned reflexes, 2014; Bergh et al., 2014). Early death also results as shown by professional footballers and Olympic when normal growth and development is followed track and field champions; then a more rapid by adverse circumstances which accelerate the rate decline to morbidity and eventual death (full line in of functional loss (dashed line in figure 2); for figure 2). example, when health selection into employment is Social and material circumstances can affect both followed by years of physically arduous and the rate of growth and development and the rate of hazardous work (Costa & d’Errico, 2006). The social functional decline. Adverse circumstances can structure, with its tendency towards life course constrain growth and development to produce sub- continuity of advantage or disadvantage optimal peak functioning, with the result that (Goldthorpe, Llewellyn & Payne, 1980; Berney et al., subsequently the normal rate of loss of function will 2000; Holland et al., 2000; Erikson & Goldthorpe, still produce early death (dotted line in figure 2); for 1993), ensures that some people experience both example, in Swedish linked register data childhood sub-optimal growth and accelerated decline. social disadvantage, in terms of parental

168

COMMENT AND DEBATE

Applying figure 2 to a UK citizen currently aged Work and fertility 1949-1988/1993 85 years shows that their period of growth and Male members of the cohort ended their phase development would have occurred during 1928- of growth and development conscripted into 1948 and their subsequent decline during 1949- compulsory national service in UK armed forces, 2013, which latter period can be sub-divided with its physical fitness, calorie-rich diet and heavily usefully into working life (women 1949-1988; men subsidised tobacco smoking. Female members of 1949-1993) and retirement (women 1989-2013; the cohort worked in light industry rather than men 1994-2013). The UK social history context in domestic service, as formerly, and married in which each of these stages was lived will be unprecedented numbers, producing fewer children examined next. despite rudimentary contraception (Coleman, Growth and development 1928-48 2000). The cohort raised their children in residential circumstances that, initially, were poor, often This golden cohort’s childhood was spent during crowded, unhygienic and cold. That improved the economic of the 1930s and their slowly during the following two decades, by which adolescence during World War II and its post-war time bathrooms, indoor toilets, separate bedrooms austerity. What this meant in terms of childhood for adults and adolescents, refrigerators and social circumstances varied by geographic region kitchen cookers were becoming universal, with and social class. New light industries in the east telephones and washing machines already midlands and south-east England brought widespread (Office for Population Censuses & employment and less hazardous work to some of Surveys, 1973). the cohort’s parents, although pay remained low The slow but steady improvement in the and jobs insecure, together with new suburban residential circumstances of the cohort was funded housing and commuting via new transport by full employment and a labour shortage that infrastructure. The heavy industrial areas of north drove rising real wages and immigration from and west Britain, in contrast, experienced high mainland Europe, Ireland, the Caribbean and the unemployment, protracted poverty and poor Indian sub-continent. The rise in real wages was housing (Stevenson, 1984). The high prevalence of accompanied by the abolition of Saturday morning poor diets among workers’ children in the 1930s working as part of the standard working week and was documented by John Boyd Orr’s careful studies the spread of paid holidays. The combination of of child health and nutrition (Harvey, 1955; Gunnell, better-equipped kitchens, new ideas from holiday 1996; Maynard et al., 2006) and motivated experience and more disposable income made demands for welfare payments (family allowances) possible better nutrition. The later years of the to cover some of the extra costs of children cohort’s working lives, when the risk of the onset of (Rathbone, 1924). chronic disease increases, coincided with de- The war brought family disruption to the cohort’s industrialisation and the 1980s economic recession adolescence, as the result of the conscription of – although the cohort was protected to some fathers and older siblings into the armed forces and extent by long-service priority access to less the evacuation of urban children to rural areas, physically demanding jobs and, for those who lost while bombing produced casualties and widespread their jobs, by pension schemes which allowed early damage to the housing stock and basic utilities like retirement on grounds of poor health and, for sewage. More positively, war-time full employment manual workers, benefits rather than brought wages to mothers, which enabled them to unemployment (Schuller, 1987). buy their somewhat meagre rations of food, and jobs to the adolescents when they left education Retirement 1989/1994-2013 (Stevenson, 1984). The cohort completed its period The cohort’s retirement from paid employment of growth and development in post-war austerity, coincided with the emergence of the idea of the continued food rationing and a severe housing Third Age as a new stage of the life course situated shortage. At the same time welfare state reforms between the end of responsibility for work and gave hope for the future and, perhaps more children and the onset of physical dependency; importantly, the cohort was established in the functionally healthy and in receipt of occupational labour market (Marwick, 1982). or private second pensions, such Third Agers are seen as free to pursue their own self-realisation and

169

COMMENT AND DEBATE pleasure (Laslett, 1989). Morris’s research into the Job insecurity, change and loss minimum income for healthy living for retired The cohort mostly entered the labour market people (Morris, Wilkinson, Dangour, Deeming & during the full employment of World War II and Fletcher, 2007) shows that many in the cohort will benefitted from the later labour shortage in terms have been denied such a Third Age, particularly of secure and rising incomes and the ease of finding those afflicted by chronic disease (Patsios, 2014). alternative employment. However the existence of the idea indicates that the cohort’s retirement is being lived in new Onset of chronic illness and labour market exit circumstances, which includes new The cohort reached the age where chronic about early in terms of the nutrition and disease starts to become more prevalent during the physical exercise required to maintain functioning high unemployment and de-industrialisation of the and the ability of social participation to confer 1980s, when premature labour market exit was resilience to the adversities of . possible through early retirement on an occupational pension or, for those without an occupational pension, through permanent sickness Social Life Course on disability benefit. The life course is a social as well as a biological In each case, those from an advantaged social phenomenon, so the foregoing biological class were more likely to complete successfully the perspective needs to be complemented by its social transition faced, but for the cohort as a whole the equivalent. later challenges occurred in more favourable Bartley, Blane and Montgomery (1997) identified a circumstances than the earlier ones. series of key social transitions that are faced by most people as part of current social organisation; such transitions can set a person on a long-term Discussion trajectory towards advantage or disadvantage, The biological and social life courses point in the depending on whether the transition is completed same direction: namely, that those currently aged successfully. The sites of such transitions include 85 years in UK mostly did not have a particularly education, family, work and ageing; and each needs advantaged childhood and adolescence, while their to be seen in its cohort-specific social and material adulthood and early old age were marked by context. steadily improving social and financial circumstances. These findings suggest that life The move from primary to secondary school course research might question its emphasis on and school examinations early life, give more attention to adult This transition would have had little relevance circumstances and look at the continuities between for those currently aged 85 years because they these stages of life. reached the then minimum school leaving age of 14 At least three objections to this line of reasoning years in 1942, when perhaps 86 per cent left school need to be addressed. First, can these ideas be with no qualifications (Stevenson, 1984). University tested; can positive health be defined more study was rare among this cohort: 2.6 per cent at precisely and can it be measured? Second, the the 1951 decennial census (Carr-Saunders, Caradog increase in life expectancy at middle age has Jones & Moser, 1958) and disrupted by national continued beyond the golden cohort, in conscription in UK armed forces. circumstances where the balance between early and later life may be different. Third, the increase in Leaving parental home, establishing own life expectancy at middle age is a worldwide home, parenthood phenomenon; can a worldwide phenomenon be By the mid-1950s when the cohort was making explained by events in one country (UK)? the transition to their own home and family, the Each of these objections suggests a new line of necessary residential accommodation was in short enquiry. The measurement of some aspects of supply and often lacked basic amenities. As a result, positive health is long established, particularly in many families were started in an in-law’s home, in relation to the phase of growth and development. consequently crowded circumstances, lacking Formerly in the UK, the school medical service privacy. measured height and weight during childhood; the

170

COMMENT AND DEBATE national educational test measured cognitive where the proportionate fall in mortality from the function at age 11 years; and the army medical prevalent causes of death differed from that in corps measured various aspects of the physical and England & Wales; and if so, what were the psychological functioning of those conscripted into for this difference? Are there countries where the the armed forces at the end of adolescence. Apart increase in life expectancy at middle age started from special surveys of population samples, earlier or later than in Britain; and if so, do they comparable measures during the phase of show the same gender differences? Are there functional decline are rare, particularly during its countries whose history reversed the UK life course early stages, before overt morbidity; routine sequence of childhood hardship followed by screening by occupational health departments and growing adult affluence; and if so, what was the private health insurance companies are the main associated change in life expectancy at middle age? exceptions. Most research has concentrated on Once again, these research questions are timely specific measures such as coronary heart disease because of the creation of international risk, although measures of positive health more demographic databases, such as that of Max Planck generally have been proposed, such as the concept Institute in Rostock, collaborating with Rand of allostatic load (McEwen & Seeman, 1999) and its Corporation in USA, and international bio-social current development (Kelly-Irving et al., 2014). surveys, such as the USA National Institute of Aging New initiatives in the area of positive health would portfolio studies of ageing. be timely because of the wealth of biomedical data In summary, the present discussion paper has becoming available in large, representative social argued that: studies of longevity risk could usefully surveys, such as the Constances cohort (Zins et al., include the idea of positive health; that the life 2010) and the UK Household course perspective offers a way of doing so which is (Benzeval, Devillas, Kumari & Lynn, 2014). biologically and socially plausible; that it is The second and third objections require instructive to set these life course stages in their international comparative research to identify cohort-specific social history context; and that countries whose demographic and social history doing so suggests a timely and progressive differs from those of the UK. Are there countries programme of research.

Acknowledgements We owe a debt of gratitude to Madhavi Bajekal, Rosalind Raine and the members of the Institute and Faculty of Actuaries, particularly Joseph Lu, Brian Ridsdale and Richard Willetts, for raising our interest in longevity risk; and to UK Economic and Social Research Council who funded (Award No: ES-J019119/1) the International Centre for Life Course Studies in Society and Health (ICLS) of whose scientific programme the present paper is a part.

References Akinwale, B., Lynch, K., Wiggins, R., Harding, S., Bartley, M., & Blane, D. (2011). Work, permanent sickness and mortality risk: a prospective of England and Wales, 1971-2006. Journal of Epidemiology and Community Health, ;65, 786-792. http://dx.doi.org/10.1136/jech.2009.099325 Bartley, M., Blane, D., & Montgomery, S. (1997). Health and the life course: why safety nets matter. British Medical Journal, 314, 1194-1196. http://dx.doi.org/10.1136/bmj.314.7088.1194 Bajekal, M., Scholes, S., O’Flaherty, M., Raine, R., Norman, P., & Capewell, S. (2013). Unequal trends in coronary heart disease mortality by socioeconomic circumstances, England & Wales 1982-2006. PLoS One,8:e59608. http://dx.doi.org/10.1371/journal.pone.0059608 Benzeval, M., Devillas, A., Kumari, M., & Lynn, P. (2014). Understanding Society – the UK Household Longitudinal Study: biomarker user guide & glossary. University of Essex, Institute of Social and Economic Research.

171

COMMENT AND DEBATE

Bergh, C., Udumyan, R., Fall, K., Nilsagard, Y., Appelros, P., & Montgomery, S. (2014). Stress resilience in male adolescents and subsequent risk: cohort study. Journal of Neurology, Neurosurgery and Psychiatry, 85, 1331-1336. http://dx.doi.org/10.1136/jnnp-2013-307485 Berney, L., Blane, D., Davey Smith, G., Gunnell, D., Holland, P., & Montgomery, S. (2000). Socioeconomic measures in early old age as indicators of previous lifetime exposure to environmental health hazards. of Health and Illness, 22, 415-430. http://dx.doi.org/10.1111/1467-9566.00212 Bianconi, E., Piovesan, A., Facchin, F., Beraudi, A., Casadei, R., Frabetti, F., Vitale, L., Pelleri, M.C., Tassani, S., Piva, F., Perez-Amodio, S., Strippoli, P. & Canaider, S. (2013). An estimation of the number of cells in the human body. Annals of Human Biology, 40(6), 463-471. http://dx.doi.org/10.3109/03014460.2013.807878 Blane, D., Kelly-Irving, M., d’Errico, A., Bartley, M., & Montgomery, S. (2013). Social-biological transitions: how does the social become biological? Longitudinal and Life Course Studies, 4, 136-146. Carr-Saunders, A., Caradog Jones, D., & Moser, C. (1958). A survey of social conditions in England and Wale (p. 58). Oxford: Clarendon Press. Coleman, D. (2000). Population and family. In A. Halsey, & J. Webb (Eds.), Twentieth century British social trends (pp.27-93). London: Macmillan. Costa, G., & d’Errico, A. (2006). Inequalities in health: do occupational risks matter? European Journal of 26 (2). http://dx.doi.org/10.1093/eurpub/ckl112 Erikson, R., & Goldthorpe, J. (1993). The constant flux. Oxford: Clarendon Press. Goldthorpe, J., Llewellyn, C., & Payne, C. (1980). Social mobility and class structure in modern Britain. Oxford: Clarendon Press. Gunnell, D. (1996). Food, death and Income – a follow-up study based on a survey of family diet and health in pre-war Britain (1937-39). University of Bristol, Doctoral Thesis. Harvey, D. (1955). Family diet and health in pre-war Britain: dietary and clinical survey (Report to the Carnegie United Kingdom Trust from the Rowett Research Institute). Dunfermline: Carnegie United Kingdom Trust. Holland, P., Berney, L., Blane, D., Davey Smith, G., Gunnell, D., & Montgomery, S. (2000) Life course accumulation of disadvantage. Social Science and Medicine, 50, 1285-1295. http://dx.doi.org/10.1016/S0277-9536(99)00375-5 Human Mortality Database website. Data accessed as follows. Data: Zipped Data Files; Period Data: Death Rates; enter user name & password; select Mx 1x10; select GBRTENW. Kelly-Irving, M., Barbosa-Solis, C., Fantin, R., Lamy, S., Lang, T., & Delpierre, C. (2014). Adverse childhood experiences and allostatic load in mid-life: pathways between early life experience and adult health decline. European Journal of Public Health, 26 (1). Laslett, P. (1989). A fresh map of life: the emergence of the Third Age. London: Weidenfeld and Nicolson. Marwick, A. (1982). British society since 1945. London: Penguin. Maynard, M., Gunnell, D., Abraham, L., Ness, A., Bates, C., & Blane, D. (2006). What influences diet at older ages? Prospective and cross-sectional analyses of the Boyd Orr cohort. European Journal of Public Health, 16, 315-323. http://dx.doi.org/10.1093/eurpub/cki167 McEwen, B., & Seeman, T. (1999). Protective and damaging effects of mediators of stress: elaborating and testing the concepts of allostasis and allostatic load. Annals of the New York Academy of Sciences, 896, 30-47. http://dx.doi.org/10.1111/j.1749-6632.1999.tb08103.x McKeown, T., & Lowe, C. (1966). An introduction to social medicine. Oxford: Blackwell. McKeown, T. (1979). The role of medicine. Oxford: Blackwell. Morris, J.N., Wilkinson, P., Dangour, A., Deeming, C., & Fletcher, A. (2007) Defining a minimum income for healthy living (MIHL): older age, England. International Journal of Epidemiology, 36, 1300-1307. http://dx.doi.org/10.1093/ije/dym129 Office for National Statistics website. Data accessed as follows: Themes: Population. Deaths: Mortality Rates. Vital Statistics: Population and Health Reference Tables. select current & past editions. Spring 2011. select Reference Tables. open Zip file. select Table 6.1 xls.

172

COMMENT AND DEBATE

Office for Population Censuses and Surveys (1973). General household survey 1971 (and subsequently each year). London: HMSO. Patsios, D. (2014). Pensioner poverty and social exclusion. www.poverty.ac.uk. Raine, R., Wong, W., Ambler, G., Hardoon, S., Petersen, I., Morris, R., Bartley, M., & Blane, D. (2009). Socio- demographic variations in the contribution of secondary drug prevention to stroke survival at middle and older ages: cohort study. British Medical Journal, 338, b1279. http://dx.doi.org/10.1136/bmj.b1279 Rathbone, E. (1924). The disinherited family. London: Edward Arnold. Schuller, T. (1987). Second adolescence? The transition from paid employment. Work, Employment and Society, 1, 352-370. http://dx.doi.org/10.1177/0950017087001003005 Strachan, D., & Sheikh, A. (2004). A life course approach to respiratory and allergic diseases. In D. Kuh, & Y. Ben-Shlomo (Eds.) A life course approach to chronic disease epidemiology (2nd Edition) (pp.240-259). Oxford: Oxford University Press.. Stevenson, J. (1984). British society 1914-45. London: Penguin. Sundin, P., Udumyan, R., Sjostrom, P., & Montgomery, S. (2014). Predictors in adolescence of ESRD in middle-aged men. American Journal of Kidney Disease, 64, 723-729. http://dx.doi.org/10.1053/j.ajkd.2014.06.019 Unal, B., Critchley, J., & Capewell, S. (2005). Modelling the decline in coronary heart disease deaths in England & Wales, 1981-2000. British Medical Journal, 331, 614-615. http://dx.doi.org/10.1136/bmj.38561.633345.8F Willets, R. (2004). The cohort effect: insights & explanations. Cambridge: Cambridge University Press Willets, R., Gallop, A., Leandro, P., Lu, J., Macdonald, A., Miller, K., Richards, S., Robjohns, N., Ryan, J., & Waters, H. (2004). Longevity in the 21st century. London: Institute of Actuaries & Faculty of Actuaries. World Health Organisation (2008). : Closing the gap in a generation. Commission on Social Determinants of Health, Geneva: World Health Organisation. Zins, M., Bonenfant, S., Carton, M., Coeuret-Pellicer, M., Gueguen, A., Gourmelen, J., Nachtigal, M., Ozguler, A., Quesnot, A., Ribet, C., Rodrigues, G., Serrano, A., Sitta, R., Brigand, A., Henny, J., & Goldberg, M. (2010). The CONSTANCES cohort: an open epidemiological laboratory. Public Health, 10, 479-487. http://dx.doi.org/10.1186/1471-2458-10-479

173

COMMENT AND DEBATE

Commentary by Hans-Werner Wahl Heidelberg University, Germany [email protected]

Life-span and social-behavioural aging science: need for better liaisons with life course epidemiology in the future

Life-span developmental psychology and life London in 1885. Life-span oriented research also course research in the sociology of ageing (see received a strong push by Austrian developmental below for differences) have been established meta- psychologist Charlotte Bühler’s (1933) widely perspectives in social-behavioural ageing science for acknowledged monograph ‘Der menschliche some decades. However, such ‘old’ meta- Lebenslauf als psychologisches Problem’ (‘Human perspectives face permanent challenges to re- Life Course as a Psychological Problem’). invent themselves as a result of the continuing More recently, particularly after ’s development of new theories and the emergence of instalment as a science shortly after World War II, new longitudinal data, as well as synergies arising two separated traditions, operating with different from the interchange with other disciplinary views. I names, have evolved in the sociology and the will use the Blane, Akinwale, Landy, Matthews, and psychology of ageing. In sociology, the term life Wahrendorf article (2016) as a springboard to take course has mostly been used and a central a closer look at what life-span developmental argument always has been that social forces largely psychology and life course research has achieved so determine the flow of life and define expectations far, particularly in empirical but also in conceptual about what should happen in a ‘normal’ life at a terms. At the same time, I will refer to this material certain age (e.g., Settersten, 2003). Psychology to take a critical look at the Blane et al. (2016) work. preferred the term life-span development and predominantly concentrated on proximal and distal A glance at the history of life-span / influences of a range of factors such as personality, life course views in social and cognitive functioning, or childhood trauma, in order to understand the dynamics of life-long, behavioural ageing research developmentally relevant processes as well as to The insight that it might not be a good idea to predict late-life outcomes (e.g., Baltes et al., 2006). split the human life course and concentrate, for These two lines of research converged at times, but example, only on early childhood or on old age is also searched to define themselves from each other not new in developmental science. Groffman (1970) across the decades (e.g., life-span researchers also has provided a comprehensive account of European use the cohort concept to understand historical historical roots of life-span thinking in philosophy, changes in late-life cognitive functioning; sociologist culture and fictional literature. Examples of have a tendency to accuse life-span researchers of influential philosophical perspectives on ageing under-rating the role of (SES) include: Francis Bacon in England (1561-1626; across full lives). More recently, health research has ‘History of Life and Death’), Michel de Montaigne significantly added to both life course as well as life- (1533-1592; ‘Life Wisdom’s Final End’) and span research and it seems that epidemiology Johannes Nikolaus Tetens in Germany (1736 – 1807; increasingly searches for the inclusion of distal ‘Philosophical Treatise on Human Nature and its antecedent factors, when it comes to the Development’). One should also mention Belgian predictions of a range of endpoints such as diseases mathematician and statistician Adolphe Quetelet and functional limitations (e.g., Ben-Shlomo & Kuh, (1835), who argued for the need to detect the laws 2002; Kuh et al., 2014). of life-long development in his two-volume treatise In conclusion, a life course / life-span view has ‘Sur l'homme et le développement de ses facultés’ existed for some time, when it comes to social- (‘On man and the development of his capabilities’), behavioural research on ageing and, to some and Francis Galton from England, who undertook extent, also in the area of health and aging one of the first cognitive performance studies research. I therefore find the framework proposed across a wide age span – from infancy to old age – by Blane et al. to be less innovative when seen from during the International Health Exhibition held in

174

COMMENT AND DEBATE the perspective of social-behavioural aging science. research also underscores significant interactions I was also surprised that the relationship to the life between expression of specific genes and early-life course approach in the health and epidemiology traumatic experiences, leading to different mental area, e.g., as suggested by Kuh and colleagues (e.g., health outcomes including abnormal behaviours, Kuh et al., 2014), was not described. which again may dramatically shape the remaining life course. Going further, Schafer and Ferraro Important data and findings of life-span (2012) have shown with data from the US targeting and life course research: a selective and early-life conditions based on rather objectively assessed retrospective information, that early-life exemplary view with a focus on health- economic conditions do show substantial related outcomes associations with successful versus less successful Setting aside the more historically relevant (‘positive health’) lifespan developmental distinction between life-span and life course trajectories including wellbeing in old age. Brandt et research, a number of findings with importance for al. (2012) show similar findings based on data from late-life health issues emerged in social-behavioural a number of European countries. Emerging findings ageing research and some of the prototypical ones from other research teams also found similar will be mentioned in what follows. First, in the relationships between adverse childhood contexts emerging area of the critical role that differences in and late-life functional ability impairment, cognitive functioning seem to play for late-life multimorbidity, and mortality (Pavela & Latham, outcomes (‘cognitive epidemiology’; see Calvin et 2015; van den Berg et al., 2006, 2009). Most of al., 2011), work by researcher Deary at these associations remained statistically meaningful al. has shown that cognitive functioning as even after controlling for proximal health and SES measured in 11-year-old children reveals a very conditions. strong relationship with intelligence assessed 70 Fourth, the existence of early health risk factors years later (Deary et al., 2004). In addition, Deary seems to impact on the full rest of the life span. A and colleagues (2004) also found, in line with a relatively large and fairly consistent body of number of other studies (see also Schaefer et al., evidence from epidemiological studies now 2015), that being higher in intelligence in childhood demonstrates that being overweight in childhood not only predicts living longer even after controlling and adolescence has adverse consequences on for a number of confounders, but also type 2 premature mortality and physical morbidity in diabetes, heart disease, and in late life. adulthood (Reilly & Kelly, 2011). Additionally, the Second, the linkage between personality and clear majority of depressive illnesses have emerged late-life health outcomes is complex but well before the age of 25 years; hence, and particularly established. In particular, being a high scorer in given the resistance to treatment of depressive neuroticism seems to predict follow-up health illness at large, late-life depression is to a large events and longevity rather well (e.g., Mroczek et extent connected with depressive episodes that al, 2006). Importantly, inter-individual differences in already appeared earlier in the life-span (Kessler et personality traits tend to be stable after young al., 2014). Similarly, low educational input, low adulthood until the end of life (Roberts & physical activity level, cognitive inactivity, midlife DelVeccio, 2000) and thus unfold their health- overweight, diabetes, smoking and depression related impact on the full remaining life course. earlier in life show relatively strong associations Third, Blane et al. already consider the critical with late-life dementia-related disorders (Barnes & role of adverse childhood conditions for very late Yaffe, 2011). health-related outcomes to some extent in the Fifth, two simple subjective evaluations assessed article. However, there is much more. Caspi and earlier in life (e.g., in midlife), i.e., subjective health colleagues in the UK have shown, in a range of and subjective age, show important relations with studies, how important early childhood experiences late-life health-related endpoints. Subjective health are for later-life outcomes via abnormal adolescent assessed in early life has been found to predict late- behaviour which in reverse may impact on the rest life morbidity and mortality even when objective of life, e.g., via depressed mood or reduced health is controlled (Jylhä, 2009). Also, feeling educational pathways (e.g., Caspi et al., 2002). His younger and having a more positive attitude

175

COMMENT AND DEBATE towards own aging, when assessed in midlife, can geriatrician and a social psychologist, have predict late-life cardio-vascular disease, functional suggested a concept of successful ageing based on ability, dementia-related physiological changes in three components: (1) low risk of disease and the brain, and mortality (e.g., Levy et al., 2002; see related disability; (2) high cognitive and physical also meta-analysis by Westerhof et al., 2014). It functioning, and; (3) continued active engagement may be argued here that feeling healthier as well as in life. All three of these components are feeling younger may represent a rather powerful substantially related to longevity and all have, as ways to resist the ageing process and thus indicate argued above, major sources in earlier life efficient forms of resilience as people age, starting conditions. They indeed also have considerable its operation rather early in the life-span. overlap with the concept of positive health as In summary, it seems rather clear that described by Blane et al. differences appearing in late-life health outcomes Second, the so-called well-being paradox has including longevity are, to a considerable extent, gained much attention in social-behavioural ageing influenced by differences earlier and even very research and we now know that ageing people have early in life – including in childhood. It is interesting a full tool box at their disposal allowing them to too that Blane et al. do not feature this body of maintain high levels of wellbeing, even in highly work, mostly coming from the social and adverse conditions. Two such mechanisms are also behavioural aging science area, very much in their echoed in the data as mentioned above, i.e. feeling article. I interpret this as an indication that life-span subjectively healthy even when faced with severe / life course approaches still operate too much in morbidity and feeling younger even beyond the age silos, with each unaware of what has already been of 80 years. In a sense, such processes may done in different disciplinary areas. This cross- resemble what Blane et al. had in mind, when they disciplinary problem needs to be overcome, referred to growth and development as well as because the concepts focused on in the empirical vitality in their conceptualisation of positive health. work outlined above are all able to speak to the Third, resilience, also explicitly addressed by concept of positive health – a core construct that Blane et al., has generated much interest in social- Blane et al. are using in their article. behavioural ageing science since the 1990’s (Staudinger & Greve, in press). As Staudinger and Concept of positive health Greve (in press) argue, resilience should be seen as The argument made by Blane et al. that a ubiquitous phenomenon of ageing and not as the traditional gero-epidemiology research targeting rare exception, although personality characteristics longevity (mortality) is too much focused on a for example (see also data mentioned above) are disease and disease-risks perspective is important. likely of importance to foster resilience. Blane et al. (2016) rightly state: “positive health, in In the light of such conceptual ideas as well as the sense of growth and development, functional the empirical work I have compiled in the previous capacity, vitality and resilience, was largely section, I feel that the term ‘positive health’ has ignored.” (p. 167). Later in their work, they also remained too vague in the Blane et al. article. We address issues of definition and assessment in the need a robust conceptual model that addresses context of positive health. Indeed, it seems that what positive health may mean and issues such as disease-specific mortality rates are on the decline (childhood) intelligence, psychological resilience, due to on-going medical progress and positive attitudes toward own aging, as well as personality health may increasingly take over as a driving force traits (e.g., low neuroticism, but also high of longevity in decades to come. conscientiousness) must all be considered in such a However, in social-behavioural ageing research, model. There is also a strong need to develop such the term positive health would be unlikely to find a model in historic-dynamic terms; hence, ongoing much resonance and the immediate association cohort changes, as just mentioned, must become would be the concept of successful ageing. Three part of such a conceptual approach. It even seems influential concepts of successful ageing may be that using ‘positive health’ is too narrow a term, distinguished. First, Rowe and Kahn (1997), a because important factors related to longevity risk seem not to belong in the health sphere.

176

COMMENT AND DEBATE

Outlook potential / risk to have ‘long-term’ impact on the Both the life course orientation as well as the remaining life course and thus become relevant for emphasis put on positive health in the work by for intervention and early preventive efforts. Blane et al., suggest that great synergies would In conclusion, the concept of positive health, unfold if life course epidemiology and life-span / life enriched and differentiated with now empirically course research in social-behavioural ageing proven concepts anchored in social and behavioural research would move closer together. First, better aging science, may further the needed Gestalt links and more cooperation between these areas Switch from a traditionally mostly disease-related may widen the range of constructs to be considered view of longevity to a view that integrates health / to better understand longevity risk. An important disease and social-behavioural parameters part of such widening would, in my view, be the important for human long-term development. In an insight that factors beyond health play a significant optimistic scenario, this may also open our eyes to role when it comes to efficient prediction models of new public health and preventive approaches that, longevity. Second, better connections may greatly to a large extent, may operate early in life, for enhance our empirical understanding of which example in kindergarten and the early school years. kinds of mechanisms are operating and linking early Hence, it may become increasingly true in the and late-life outcomes. Third, the expected future that old age outcomes as well as end of life conceptual and empirical progress based on better dynamics are linked with occurrences early in life links and more co-operation may lead to better and possibly even at the pre-natal stage of targeting of critical conditions early in life with the development.

References Baltes, P. B., Lindenberger, U., & Staudinger, U. M. (2006). Life-span theory in developmental psychology. In W. Damon & R. M. Lerner (Eds.), Handbook of child psychology (6th ed., Vol. 1: Theoretical models of human development, pp. 569-664). New York: Wiley. Blane, D., Akinwale, B., Landy, R., Mathews, K. & Wahrendorf, M. (2016). What can the life course approach contribute to an understanding of longevity risk? Longitudinal and Life Course Studies 7(2) Barnes, D. E., & Yaffe, K. (2011). The projected effect of risk factor reduction on Alzheimer’s disease prevalence. The Lancet Neurology, 10, 819-828. http://dx.doi.org/10.1016/S1474-4422(11)70072-2 Ben-Shlomo, Y., & Kuh, D. (2002). A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives. International Journal of Epidemiology, 31, 285-293. http://dx.doi.org/10.1093/ije/31.2.285 Brandt, M., Deindl, C., & Hank, K. (2012). Tracing the origins of successful aging: the role of childhood conditions and social inequality in explaining later life health. Social Science & Medicine, 74, 1418- 1425. http://dx.doi.org/10.1016/j.socscimed.2012.01.004 Bühler, C. (1933). Der menschliche Lebenslauf als psychologisches Problem [The Human Life Course a Psychological Problem]. Leipzig: Hirzel. Calvin, C. M., Deary, I. J., Fenton, C., Roberts, B. A., Leckenby, G., & Batty, G. D. (2011). Intelligence in youth and all-cause mortality: systematic review with meta-analysis. International Journal of Epidemiology, 40, 626-640. http://dx.doi.org/10.1093/ije/dyq190 Caspi, A., McClay, J., Moffitt, T., Mill, J., Martin, J., Craig, I. W., … & Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851-854. http://dx.doi.org/10.1126/science.1072290 Deary, I. J., Whiteman, M. C., Starr, J. M., Whalley, L. J., & Fox, H. C. (2004). The impact of childhood intelligence on later life: Following up the Scottish mental surveys of 1932 and 1947. Journal of Personality and , 86, 130-147. http://dx.doi.org/10.1037/0022-3514.86.1.130 Groffmann, K. J. (1970). Life-span developmental psychology in Europe: past and present. In L. R. Goulet & P. B. Baltes (Eds.), Life-span developmental psychology: Research and theory (pp. 53-68). New York: Academic Press. http://dx.doi.org/10.1016/B978-0-12-293850-4.50009-1

177

COMMENT AND DEBATE

Jylhä, M. (2009). What is self-rated health and why does it predict mortality? Towards a unified conceptual model. Social Science & Medicine, 69, 307–316. http://dx.doi.org/10.1016/j.socscimed.2009.05.013 Hülür, G., Drewelies, J., Eibich, P., Düzel, S., Demuth, I., Ghisletta, P., Steinhagen-Thiessen, E., Wagner, G. G., Lindenberger, U., & Gerstorf, D. (in press). Cohort differences in psychosocial function over 20 years: Current older adults feel less lonely and less dependent on external circumstances. Gerontology. Kessler, E.-M., Kruse, A., & Wahl, H.-W. (2014). Clinical gero-psychology: A lifespan perspective. In N. A. Pachana & K. Laidlaw (Eds.), The Oxford Handbook of Clinical Geropsychology (pp. 3-25). Oxford University Press. Kuh, D., Karunananthan, S., Bergman, H., & Cooper, R. (2014). A life-course approach to healthy ageing: maintaining physical capability. Proc Nutr Soc, 73, 237-248. http://dx.doi.org/10.1017/S0029665113003923 Levy, B. R., Slade, M. D., Kunkel, S. R., & Kasl, S. V. (2002). Longevity increased by positive self-perceptions of aging. Journal of Personality and Social Psychology, 83, 261-270. http://dx.doi.org/10.1037/0022- 3514.83.2.261 Mroczek, D. K., Spiro III, A., & Griffin, P. W. (2006). Personality and aging. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging (6th ed.; pp. 363-378). Burlington: Elsevier Academic Press. http://dx.doi.org/10.1016/B978-012101264-9/50019-7 Quêtelet, A. (1835). Sur l'homme et le développement de ses facultés [A Treatise on Man and the Development of his Faculties]. Paris: Bachelier. Pavela, G., & Latham, K. (2015). Childhood conditions and multimorbidity among older adults. Journals of Gerontology: Psychological Sciences and Social Sciences. Online published. http://dx.doi.org/10.1093/geronb/gbv028 Reilly, J. J., & Kelly, J. (2011). Long-term impact of overweight and in childhood and adolescence on morbidity and premature mortality in adulthood: systemic review. International Journal of Obesity, 35, 891-898. http://dx.doi.org/10.1038/ijo.2010.222 Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126, 3-25. http://dx.doi.org/10.1037/0033-2909.126.1.3 Rowe, J. W., & Kahn, R. L. (1997). Successful aging. The Gerontologist, 37, 433-440. http://dx.doi.org/10.1093/geront/37.4.433 Schaefer, J. D., Caspi, A., Belsky, D. W., Harrington, H., Houts, R., Israel, S. …Moffitt, T. E., (2015). Early-life intelligence predicts midlife biological age. Journal of Gerontology: Psychological Sciences and Social Sciences. Online Published. http://dx.doi.org/10.1093/geronb/gbv035 Schafer, M. H., & Ferraro, K. F. (2012). Childhood misfortune as a threat to successful aging: Avoiding disease. The Gerontologist, 52, 111-120. http://dx.doi.org/10.1093/geront/gnr071 Settersten, R. A. (2003). Propositions and controversies in life-course scholarship. In R. A. Settersten (Ed.), Invitation to the life course: toward new understandings of later life (pp. 15-45). Amityville, NY: Baywood. Staudinger, U. M., & Greve, W. (in press). Resilience and aging. In N. Pachana (Ed.), Encyclopedia of geropsychology. New York: Springer. http://dx.doi.org/10.1007/978-981-287-080-3_122-1 Van den Berg, G.J., Lindeboom, M. & Portrait, F. (2006). Economic conditions early in life and individual mortality. American Economic Review, 96, 290-302. http://dx.doi.org/10.1257/000282806776157740 Van den Berg, G. J., Doblhammer, G., & Christensen, K. (2009). Exogenous determinants of early-life conditions, and mortality later in life. Social Science & Medicine, 68, 1591-1598. http://dx.doi.org/10.1016/j.socscimed.2009.02.007 Westerhof, G., Miche, M., Bothers, A., Barrett, A., Diehl, M., Montepare, J., Wahl, H.-W. & Wurm, S., (2014). The influence of subjective aging on health and longevity: A meta-analysis of longitudinal data. Psychology and Aging, 29, 793-802. http://dx.doi.org/10.1037/a0038016

178

COMMENT AND DEBATE

Commentary by Mark D. Hayward University of Texas at Austin, USA [email protected]

The paper by Blane, Akinwale, Landy, Matthews and less exposure to lifetime health risks than and Wahrendorf provides a thoughtful and earlier cohorts, contributing in important ways to nuanced discussion of the value of the life course historical declines in US adult mortality (Yang, framework for understanding the origins of adult 2008). Thus, the health of birth cohorts, often health and mortality. Most highly-prevalent adult reflected in age-specific trends in health, may health problems in the UK and other high income change due to fundamental shifts in the nature countries are a long time in the making and reflect and timing of life course experiences. A key issue exposures stretching from the prenatal that is implicit in the discussion by Blane and environment and childhood, to adolescence, and colleagues is that life course influences on adult adulthood. Because the social conditions, health are largely endogenous to the historical institutions, and stratification systems of societies context. I suggest that we are only at the initial are powerful forces that shape the nature of stages of understanding how life course influences individuals’ life courses, adult health is on health are being transformed by changes in the fundamentally a reflection of these lifetime social social capacity for health. Their call for forces. comparative research is an important component As Blane and his colleagues observe in their in furthering this agenda. review of the historical conditions surrounding the Despite growing agreement over this life course of the UK’s golden cohort, social conceptual framework, Blane and his colleagues conditions, institutions and stratification systems alert us to the fact that applying this framework to are far from static. Over the course of the 20th adult health outcomes comes with important century, for example, the social capacity for health challenges. Three major challenges that I discuss in many countries, particularly high income below are the following. First, how ought countries, has dramatically improved through researchers to define health? I argue that we technological innovation, dramatic growth in should consider a ‘portfolio’ of health outcomes biomedical knowledge and improvements in social encompassing biological risk, morbidity, institutional resources (Easterlin, 1997). These functioning and disability, and mortality. A second societal forces have had rippling consequences for challenge is that conceptual frameworks of social changes in the nature and timing of life course factors influencing health are often not biologically experiences for cohorts born at different points in informed. I agree with Blane and colleagues that it history. In turn, changes in the nature and timing is critical to integrate biological and social life of life course experiences have led to a host of course frameworks in order to understand how life changes in population health. Conditions once course exposures from childhood into adulthood important in defining the health of the population shape health trends and disparities through both have waned (e.g., the long-run decline in heart developmental and aging processes. Finally, Blane disease or the near eradication of polio in most and colleagues argue for the importance of high income countries), while other conditions incorporating historical context in understanding have become more prevalent (e.g., some ). trends and differences in the life course pathways The concept of the birth cohort is critical in leading to adult health problems. Dramatic understanding these trends, due to cohort changes have occurred across current birth cohorts differences in the nature and timing of exposures. represented in the adult population in high-income For example, given the role of childhood countries in their prenatal, childhood and adult vaccinations in improving child health and survival exposures, yet these changes are rarely central in in the 20th century in the United States (Centers for life course studies of health. The development of Disease Control and Prevention, 1999; Andre et al., conceptual models requires sensitivity to the fact 2008), combined with the dramatic decline in adult that stratification systems, institutions, social smoking (Fenelon and Preston, 2012), more recent conditions, technology, and even the American birth cohorts have experienced fewer epidemiological environment are changing and

179

COMMENT AND DEBATE differ across countries, all of which have results in a cascade of increased risks of functional implications for tackling a comparative research problems and mortality from cardiovascular agenda. disease (CVD). In this hypothetical example, childhood health problems’ association with CVD- Challenge one related mortality stems primarily from increasing Only recently have researchers begun to take a disease incidence. In contrast, other life course more integrative view of adult health and factors may come into play throughout the examined how life course experiences influence process. Educational attainment, for example, not multiple facets of health (e.g., how life course only is associated with a lower risk of heart disease factors influence the interplay of functioning and but potentially it also comes into play in mortality to determine healthy life expectancy differentiating the risks of functional problems and (Montez & Hayward, 2014)). Research rarely mortality among persons with heart disease explicitly addresses the idea that exposures may through human agency and financial resources give rise to a cascade of adult health conditions (Mirowsky & Ross, 2003). In this example, a major starting, say, from morbidity, to disability, and social resource acquired relatively early in life – then to mortality – or not. All too often, research educational attainment – may accentuate assumes that such a cascade exists, it gives rise to disparities throughout the process of health more ‘endogenous’ health outcomes such as change (Manton, Stallard & Corder, 1997; Merkin, disability and mortality, and is unidirectional – Karlamangla, Crimmins, Hayward, & Seeman, despite strong cautions about these assumptions 2009). (Verbrugge & Jette, 1994). Health at the population level is a Challenge two multidimensional concept. Increasingly, definitions Blane and colleagues offer a biological, of health acknowledge the core domains of developmental framework of life course health physiological dysregulation (e.g., metabolic (see figure 2 in their article). The core idea is that functioning), conditions (e.g., disease conditions processes of the major biological systems (e.g., such as diabetes), functioning (e.g., physical and endocrine, immune, neurological, respiratory) cognitive deficits) and important facets of display a similar pattern of development – a steady wellbeing and health potential (e.g., ability to live curvilinear growth in functional capacity during independently). The measurement of these childhood followed by maintenance and eventually domains is complex as are the relationships among some decline in adulthood (Halfon & Hochstein, these domains. Physiological dysregulation, 2002). Figure 2’s conceptual framework conveys a disease conditions, functioning loss, and frailty are number of important ideas to keep in mind when all parts of the process of health change that can – considering the factors that come into play in but need not -- precede death (Crimmins, Kim & understanding differences/disparities in adult Vasunilashorn, 2010; Martin, Schoeni & Andreski, health. First, differences in capacity begin as early 2010; Crimmins & Beltrán-Sánchez, 2011). The as in utero and are evident at birth. This speaks to importance of understanding these connections the seminal work by Barker and his colleagues stems from the fact that trends in these domains (Barker, 1997; Barker, 1998; Barker, 1999; Barker, need not move in the same direction, and social 2004). Second, childhood is a period of growth in group differences may vary, depending on the capacity, and differences in capacity widen domain of interest. because of differences in inputs throughout Given these associations between health childhood (e.g., SES, diet, family relationships). This domains, life course factors need not affect all is also the period in which cohort morbidity parts of the process in the same way. Some life phenotypes (e.g., lifelong health risks that accrue course influences may be germane only to certain through early life exposures such as infection and parts of the process of health change. For example, inflammation) are established (Finch & Crimmins, it is plausible that some childhood health problems 2004; Crimmins & Finch, 2006). Although Blane (e.g., infectious conditions that heighten and colleagues note that the maximum level of inflammation) give rise to adult morbidity growth is a product of the level of inputs, I would conditions (e.g., coronary heart disease) that add that the period of growth (and different peak

180

COMMENT AND DEBATE ages of capacity) might also be highly dependent course influences on adult health has occurred in on the level of inputs. In this sense, both the length the context of single birth cohorts or a relatively and degree of development are highly malleable to narrow band of birth cohorts, e.g., the UK golden inputs (Finch & Crimmins, 2004). Third, declines in cohort described by Blane and colleagues, the capacity can start at different ages and exhibit British cohort studies and the US Health and varying rates of decline. Resources such as quality Retirement Study. Thus, much of the empirical jobs, marriages and the avoidance of disease result evidence about the life course origins of adult in slow rates of decline, while the lack of resources health is necessarily framed by the experiences of and risk factors (e.g., smoking) contribute to individuals in these studies. With the explosion of relatively accelerated rates of decline. Indeed, it is international studies built on the life course possible that despite early life developmental framework of the Health and Retirement Study advantages, adult risk factors and adverse (including the Study on Global Ageing and Health conditions can result in rates of decline that negate (SAGE)), we now have the opportunity to respond earlier advantage. The combination of lifetime to Blane and colleagues’ call for comparative gains and losses results in growing heterogeneity research – and on a global scale. of capacity in the population across most of the Blane and his colleagues make a strong case adult life course – the period of life where losses that researchers need to be highly sensitive to the occur in capacity. specific historical conditions that characterised the This biologically informed conceptual work of cohorts’ experiences at particular ages in the life life course health presented by Blane and course. Stratification systems (and their metrics) colleagues points to the need for researchers to may change across cohorts (and differ across attend to the combinations of lifetime exposures countries). For example, occupation was more that put adults at greater or lesser risk of various strongly tied to resources that garnered health health outcomes. In addition, this idea suggests advantages in the early part of the 20th century in that, depending on the inputs over a lifetime, the the US compared to educational attainment balance of childhood and adult influences on (Preston & Haines, 1991; Hayward & Gorman, health disparities can change. This is a very 2004). However, educational attainment, important idea in that the developmental ‘origins’ particularly advanced education, has grown in of adult health are highly dynamic and mutable to importance for reducing mortality in the US in the a variety of inputs across the lifetime. No phase is last half of the 20th century (Montez, Hummer, inherently more important than another in terms Hayward, Woo & Rogers, 2011; Masters, Hummer of influencing adult health disparities. That said, as & Powers, 2012; Hayward, Hummer & Sasson, Blane and colleagues aptly note, life course 2015). The 20th century in the United States also stratification processes often reinforce childhood was a period of large reductions in infectious advantages and disadvantages in adulthood, disease exposure and rising obesity. This period compounding the effects of childhood. There is no was characterised first by massive increases in ‘lottery’ after childhood that randomly assigns post-secondary educational institutions and the persons to adult trajectories of resources and risks. prevalence of post-secondary education after Moreover, some combinations of life course World War II. This trend was followed in the later exposures are likely to be relatively common while part of the century by the stalling of post- others are quite rare due to life course secondary attainment and the rise in student load stratification processes and social change. The debt. These are only a few examples but they raise malleability of life course trajectories in capacity important questions about how researchers should has significant implications for understanding measure key concepts relating to the types of social group differences and trends in the life inputs over cohorts’ lifetimes (e.g., childhood course origins of adult health – see Challenge three health, socioeconomic resources), and they point below. to some of the difficulties of carrying out a comparative research agenda. Challenge three Ryder noted 50 years ago that “the principal Not surprisingly given the constraints of motor of contemporary social change is available data, much of the research on the life technological innovation” (Ryder, 1965).

181

COMMENT AND DEBATE

Technological change is fundamentally important the social capacity for health. In addition, current for the level of social capacity for population surviving birth cohorts in high-income countries health (Easterlin, 1997). It is embedded in social are likely to differ in important ways in their institutions and defines the stock of knowledge lifetime exposures to important features of the and institutional resources that individuals in the social capacity for population health. The types of population have access to and can act on to garner exposures and the pace of change in social health advantages. The idea is similar to Fogel’s capacity necessarily will shape how birth cohorts’ concept of techno-physio (Fogel & Costa, capacities are changed over the lifetimes, 1997; Fogel, 2004) which reflects the synergistic ultimately influencing trends and disparities in the association between technological and major domains of population health. The physiological improvements in the modern era. importance of broader societal forces for health Although technological change is sometimes can thus best be appreciated through studies that thought to be felt most strongly by persons about include a range of cohorts and that analyse the to make ‘lifelong’ choices (Ryder, 1965), the data with sensitivity to between-cohort developmental trajectories shown in Blane et al.’s differences. figure 2 illustrate that technological change may have implications for changes in capacity at all Conclusions stages of the life course. The pace of technological The article by Blane and his colleagues adds to a change is an issue that is rarely considered in life burgeoning literature on the importance of life course studies of heath, yet this phenomenon has course conceptual frameworks for understanding enormous implications for understanding how the adult health and mortality. As is evident in their inputs to the growth and decline in capacity discussion, a life course approach is fundamentally ultimately influence adult health (Goldin & Katz, interdisciplinary and encompasses disciplines that 2009; Palloni & Souza, 2013; McEniry, 2014). The are key in understanding how health is shaped pace of technological change in combination with from ‘cells to society’. Epidemiology, biology, early life exposures relates strongly to the sociology, , genetics, psychology, heterogeneity of lifetime inputs to the gain and history, …, – all these fields and more are part of a loss of physical capacity shown in figure 2. growing understanding of the life course origins of An important tenet in life course conceptual adult health and mortality. The emergence of this frameworks is the importance of historical context. highly interdisciplinary approach reflects the Here, I have argued that the historical context is enormously complex web of forces by which not only relevant to a given cohort but it matters a population health is shaped and changed. great deal for the pace and nature of changes in

References Andre, F., Booy, R., Bock, H., Clemens, J., Datta, S., John, T., Lee, B., Lolekha, S., Peltola, H., Ruff, T., Santosham, M., & Schmitt, H. (2008). Vaccination greatly reduces disease, disability, death and inequity worldwide. Bulletin of the World Health Organization, 86, 140-146. http://dx.doi.org/10.2471/BLT.07.040089 Barker, D. (1999). Early growth and cardiovascular disease. Archives of Disease in Childhood, 80(4), 305. http://dx.doi.org/10.1136/adc.80.4.305 Barker, D.J.P. (1997). Maternal nutrition, fetal nutrition, and disease in later life. Nutrition, 13(9),807-813. http://dx.doi.org/10.1016/S0899-9007(97)00193-7 Barker, D.J.P.. 1998. Mothers, babies, and health in later life. London: Churchill Livingstone. Barker, D.J.P. (2004). The developmental origins of well-being. Philosophical Transactions: Biological Sciences, 359(1449),1359-1366. http://dx.doi.org/10.1098/rstb.2004.1518 Centers for Disease Control and Prevention. (1999). CDC on infectious diseases in the United States: 1900- 99. Population and Development Review, 25(3), 635-640. http://dx.doi.org/10.1111/j.1728- 4457.1999.00635.x

182

COMMENT AND DEBATE

Crimmins, E., & Finch, C. (2006). Infection, inflammation, height, and longevity. Proceedings of the National Academy of Sciences, 103(2), 498-503. http://dx.doi.org/10.1073/pnas.0501470103 Crimmins, E., Kim, J.K., & Vasunilashorn S. (2010). Biodemography: new approaches to understanding trends and differences in population health and mortality. , 47 Supplement, S41-S64. http://dx.doi.org/10.1353/dem.2010.0005 Crimmins, E.M., & Beltrán-Sánchez, H. (2011). Mortality and morbidity trends: is there compression of morbidity? The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66B(1), 75-86. http://dx.doi.org/10.1093/geronb/gbq088 Easterlin, R.A. (1997). Growth triumphant, the twenty-first century in historical perspective. Ann Arbor: The University of Michigan Press. Fenelon, A., & Preston, S. (2012). Estimating smoking-attributable mortality in the United States. Demography, 49(3), 797-818. http://dx.doi.org/10.1007/s13524-012-0108-x Finch, C.E., & Crimmins, E.M. (2004). Inflammatory exposure and historical changes in human life-spans. Science, 305(5691), 1736-1739. http://dx.doi.org/10.1126/science.1092556 Fogel, R.W. (2004). The escape from hunger and premature death, 1700-2100. New York, NY: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511817649 Fogel, R.W., & Costa, D.L. (1997). A theory of technophysio evolution, with some implications for forecasting population, health care costs, and pension costs. Demography, 34(1), 49. http://dx.doi.org/10.2307/2061659 Goldin, C.D., & Katz L.F. (2009). The race between education and technology. Boston, MA: Harvard University Press. Halfon, N., & Hochstein, M. (2002). Life course health development: an integrated framework for developing health, policy, and research. Milbank Quarterly, 80(3), 433-479. http://dx.doi.org/10.1111/1468- 0009.00019 Hayward, M.D., & Gorman, B. (2004). The long arm of childhood: the influence of early life conditions on men's mortality. Demography, 41, 87-107. http://dx.doi.org/10.1353/dem.2004.0005 Hayward, M.D., Hummer, R.A., & Sasson, I. (2015). Trends and group differences in the association between educational attainment and U.S. adult mortality: implications for understanding education's causal influence. Social Science & Medicine, 127, 8-18. http://dx.doi.org/10.1016/j.socscimed.2014.11.024 Manton, K.G., & Stallard, E., & Corder, L. (1997). Education-specific estimates of life expectancy and age- specific disability in the U.S. elderly population 1982 to 1991. Journal of Aging and Health, 9(4), 419- 450. http://dx.doi.org/10.1177/089826439700900401 Martin, L.G., Schoeni, R.F., & Andreski, P.M. (2010). Trends in health of older adults in the United States: past, present, future. Demography, 47 Supplement, S17-S40. http://dx.doi.org/10.1353/dem.2010.0003 Masters, R.K., Hummer, R.A., & Powers, D. A. (2012). Educational differences in U.S. adult mortality: a cohort perspective. American Sociological Review, 77(4), 548-572. http://dx.doi.org/10.1177/0003122412451019 McEniry, M. (2014). Early life conditions and rapid demographic changes in the developing world. Netherlands: Springer. http://dx.doi.org/10.1007/978-94-007-6979-3 Merkin, S.S., Karlamangla, A., Crimmins, E., Hayward, M., & Seeman, T. (2009). Education differentials by race in the diagnosis and management of hypercholesterolemia. International Journal of Public Health, 54(3), S1-S1. Mirowsky, J., & Ross, C.E. (2003). Education, social status, and health. New York: Aldine De Gruyter. Montez, J.K., & Hayward, M.D. (2014). Cumulative childhood adversity, education, and active life expectancy among U.S. adults. Demography, 51, 415-435. http://dx.doi.org/10.1007/s13524-013-0261-x Montez, J.K., Hummer, R.A., Hayward, M.D., Woo, H., & Rogers, R.G. (2011). Trends in the educational gradient of U.S. adult mortality from 1986 to 2006 by race, gender, and age group. Research on Aging, 33(2), 145-171. http://dx.doi.org/10.1177/0164027510392388

183

COMMENT AND DEBATE

Palloni, A., & Souza, L. (2013). The fragility of the future and the tug of the past: longevity in Latin America and the Caribbean. Demographic Research, 29(21), 543-578. http://dx.doi.org/10.4054/DemRes.2013.29.21 Preston, S.H., & Haines, M.R. (1991). Fatal years: child mortality in late nineteenth century America. Princeton: Princeton University Press. http://dx.doi.org/10.1515/9781400861897 Ryder, N.B. (1965). The cohort as a concept in the study of social change. American Sociological Review, 30(6), 843-861. http://dx.doi.org/10.2307/2090964 Verbrugge, L.M., & Jette, A.M. (1994). The disablement process. Social Science & Medicine, 38(1), 1-14. http://dx.doi.org/10.1016/0277-9536(94)90294-1 Yang, Y. (2008). Trends in U.S. adult chronic disease mortality, 1960–1999: age, period, and cohort variations. Demography, 45, 387-416. http://dx.doi.org/10.1353/dem.0.0000

184

COMMENT AND DEBATE

Commentary by Aart C. Liefbroer Netherlands Interdisciplinary Demographic Institute, University Medical Center Groningen and Vrije Universiteit Amsterdam, Netherlands

What can the life course approach contribute to an understanding of longevity risk? – a demographer’s comment

The paper by Blane et al. makes a strong case for mortality rates for all cohorts and at all ages in more attention to the life-course perspective in more recent years. Often, it will be mix of these order to better understand seminal trends in three factors. However, given that the three mortality and longevity. Given my own background elements of this classic APC (Age-Period-Cohort) in the sociology and demography of the life course, dilemma are dependent (someone’s current age it is hard to disagree with the authors on this issue, equals current period minus his or her birth cohort), so I will restrict myself to providing some brief it is hard to disentangle these elements and to reflections on a few issues of key demographic or determine whether changes in any outcome are sociological importance. influenced by age, period, cohort, or a combination The paper starts with the notion of ‘longevity of them. Statistical models are developed that try to risk’. This is terminology that ‘only’ actuaries can solve this dilemma, but without a consensus about use, as the costs to insurance companies go up if the best approach (Bell & Jones, 2014). An people live longer than expected. Most people interesting contribution from demography to would not want to view longevity mainly as a risk, understand this issue is the so-called Lexis-surface. but rather as an accomplishment. It also conflicts Recently, Minton (2014) provided a useful with the terminology of demographers, who speak illustration of its application to mortality, shown of risk (or ‘rate’) as the chance that an event will here in figure 1. It shows mortality rates of happen (rather than the event not happening), and Norwegian males, with year (period) on the X-axis view increased longevity as the outcome of a and age on the Y-axis. Birth cohort can be followed process in which mortality risks (at specific ages, for drawing a 45-degree diagonal from lower left to specific cohorts, in specific periods) have fallen. upper right. Different colours signify different levels A key issue in the Blane et al. paper is how to of mortality. Age effects are evident from the understand the strong increase in longevity. Female increase in mortality rates across the Y-axis. Period life expectancy has increased by three months per shifts are visible as mortality rates go up or down if year for more than a century (Oepen & Vaupel, one advances along the X-axis. Clear, ‘pure’ period 2002). In 2012 female life expectancy at birth was effects can, for instance, be observed during the highest in Japan at 87 years, and male life first and second world wars. A cohort shift is expectancy at birth was highest in Iceland at 82 evident, if mortality rates to the left of a cohort years (World Health Organisation, 2014). Blane et diagonal would strongly differ from those to the al. argue that this is not just the result of right. In the example, we see a strong decrease in improvements in treating specific diseases, but mortality rates from the 1930s onwards (briefly largely of general health improvements. They interspersed by World War II). The contour lines of illustrate this with period figures (the trend in all mortality levels retreat to higher ages, the closer mortality rates among 65-year-old men and we get to the current period. In the more recent women) and with a description of the golden cohort period these lines have an upward incline, (born in 1928). But it is clear that health and suggesting that they are a mix of period effects mortality differences between age groups in a (affecting virtually all age groups) and cohort effects population, or across time among the same age (with some cohorts profiting more than others). group, could depend on one of three factors. They Thus, the Lexis-surface can be a useful tool to could be age-related, with increasing mortality rates understand the dynamics of mortality decline and as people grow older. They could be cohort-related, longevity increase, which could lead to hypotheses with decreasing mortality rates among younger that could be more formally tested in a next step. cohorts. Or they could be period related with lower

185

COMMENT AND DEBATE

Figure 1. Shaded contour plot of mortality surface, Norway, males (Minton, 2014, Figure 7)

In order to understand the changes in ‘positive’ applications of the life course approach, in this health, Blane et al. emphasise the importance of respect, is the growing popularity of sequence examining the long-term consequences of earlier analysis (Abbott & Tsay, 2000; Elzinga & Liefbroer, life events for health later in life. This is not a new 2007; Aisenbrey & Fasang, 2010). Within sequence suggestion, as David Barker (1990, 1995) has analysis, a holistic approach to understanding the already alerted epidemiologists to the relevance of life course is advocated in which the effects of studying conditions during pregnancy for later occurrence, timing and sequencing of events are all development. What is important in the paper by taken into account to develop a typology of life Blane et al., is their emphasis on multiple critical course trajectories that optimally takes the periods in people’s lives that could influence their heterogeneity in people’s life courses into account. later development. This starts with early childhood The resulting typology can be used both to study where establishing secure attachment with key the precursors (including health related ones) of caregivers is of utmost importance (Bowlby, 1988). specific life course patterns, and to study the But Blane et al. suggest other critical periods later in consequences (again, including health related ones) life, such as the transition from youth to adulthood of these patterns. For the promise of a life course and retirement, as well. Fortunately, the basic ideas approach to be realised, though, it is imperative of such a life course approach already have strong that data collections within epidemiology pay more supporters within epidemiology (Power, Kuh, & attention to life course trajectories in the health Morton, 2013). A key challenge, however, is to domain, as well as to those in the family and career understand what aspects of life course change domains. For this to come about, closer matter. What are the key life domains to take into collaborations between epidemiologists and other account? Is it the experience of certain events, the health researchers on the one hand, and social time at which these events are experienced, or their scientists (in the broadest sense) on the other hand, ordering that matters? And how do events in have to be established. different life domains interact? An important recent A final plea made by Blane et al. is for more development in demographic and sociological comparative research. I could not agree more.

186

COMMENT AND DEBATE

Comparison is the essence of science. Often we SHARE (Survey of Health, Ageing and Retirement in compare social groups or categories within a Europe – a panel survey among the population aged specific spatial location (usually a specific nation 50 years and over with data being collected in 20 state). But country comparisons could help us test countries) also includes health related questions whether our results are location specific (and thus and has even moved into collecting biomarkers (see conditional on the environment) or not. In addition, www.share-project.org). These studies are designed it can alert us to factors that might explain these by social scientists, but could certainly profit from spatial differences. Unfortunately, many cross- collaboration with and involvement from national social surveys only include relatively crude epidemiologists and other health scientists. health related measures, whereas international To conclude, I fully agree with the plea by Blane medical data collections are not only very rare, but et al. for more attention to the life course paradigm often have very little information on social aspects. to better understand developments in longevity and However, there are a few exceptions (or at least in health more generally. What it takes, though, is examples to follow-up). The GGS (Generations & better data, but, above all, collaboration between Gender Survey – a panel survey among the whole specialists from different fields. Only in that way adult population with data collected in nineteen can we turn our search for understanding the social countries) includes a number of general questions underpinning of health risks into a truly on physical and mental health (see www.ggp-i.org). interdisciplinary endeavour.

References Abbott, A., & Tsay, A. (2000). Sequence analysis and optimal matching methods in sociology. Sociological Methods & Research, 29(1), 3-33. http://dx.doi.org/10.1177/0049124100029001001 Aisenbrey, S., & Fasang, A. E. (2010). New life for old ideas: the "second wave" of sequence analysis bringing the "course" back into the life course. Sociological Methods & Research, 38(3), 420-462. http://dx.doi.org/10.1177/0049124109357532 Barker, D. J. P. (1990). The fetal and infant origins of adult disease. British Medical Journal, 301, 1111. http://dx.doi.org/10.1136/bmj.301.6761.1111 Barker, D. J. P. (1995). Fetal origins of coronary heart disease. British Medical Journal, 311, 171-174. http://dx.doi.org/10.1136/bmj.311.6998.171 Bell, A., & Jones, K. (2014). Another 'futile quest'? A simulation study of Yang and Land's hierarchical age- period-cohort model. Demographic Research, 30(11), 333-360. http://dx.doi.org/10.4054/DemRes.2014.30.11 Bowlby, J. (1988). A secure base: parent-child attachment and healthy human development. New York, NY, US: Basic Books. Elzinga, C. H., & Liefbroer, A. C. (2007). De-standardization of family-life trajectories of young adults: a cross- national comparison using sequence analysis. European Journal of Population, 23(3-4), 225-250. http://dx.doi.org/10.1007/s10680-007-9133-7 Minton, J. (2014). Real geographies and virtual landscapes: exploring the influence on place and space on mortality Lexis surfaces using shaded contour maps. Spatial and Spatio-temporal Epidemiology, 10, 49-66. http://dx.doi.org/10.1016/j.sste.2014.04.003 Oepen, J., & Vaupel, J. W. (2002). Broken limits to life expectancy. Science, 296, 1029-1031. http://dx.doi.org/10.1126/science.1069675 Power, C., Kuh, D., & Morton, S. (2013). From developmental origins of adult disease to life course research on adult disease and aging: insights from birth cohort studies. Annual Review of Public Health, 34(1), 7-28. http://dx.doi.org/10.1146/annurev-publhealth-031912-114423 World Health Organisation (2014). World health statistics. Geneva: World Health Organisation.

187

COMMENT AND DEBATE

Commentary by Gita D. Mishra University of Queensland, Australia [email protected] Isabel Ferreira University of Queensland, Australia Ilona Koupil Stockholm University Karolinska Institute, Sweden

Longevity risk describes the phenomenon of adult life” (Kuh, Ben-Shlomo, Lynch, Hallqvist & living longer than predicted (Willets et al., 2004). As Power, 2003). Its aim “is to elucidate the biological, Blane and colleagues explain, the concept was behavioural, and psychosocial processes that developed by the actuarial profession as a response operate across an individual’s life course, or across to the unexpected decline in middle aged mortality generations, to influence the development of during the late 20th century that placed sudden disease risk” (Kuh et al,, 2003). Thus at the heart of financial strain on insurance and pension funds in life course research there is an emphasis on the UK and elsewhere. It may also illustrate what is statistical models that reflect basic hypotheses on known as the black swan event: a surprise, with a the role of timing or duration of exposures (Mishra major impact, that is rationalised by hindsight as if et al., 2009; Mishra, Chiesa, Goodman, De Stavola & it had been expected (Taleb, 2007). Blane and Koupil, 2013). colleagues posit that to find the key driving factor Blane and colleagues draw on two examples of for this increased life expectancy at middle age, one evidence to support the additional role of positive needs to look beyond changes in disease risk health, beyond that just from a reduction in risk factors, morbidity, and medical innovations to factors and medical innovations. Here we focus on improvements in positive health in the sense of their second example: time series data for the “growth and development, functional capacity, mortality at age 65 years in the UK that shows a vitality and resilience”. Second, they then draw on near linear decline for women since 1900. By the UK social history of a golden cohort, men and comparison, mortality for men at age 65 years is women currently aged 85 years, to argue that a life higher than for women but has a slightly larger course approach is needed to identify plausible overall decline over the same time period. It biological and social pathways that lead to these includes a plateau in mortality rates for the three changes in positive health. decades after about 1930 that were followed by a We agree that positive health appears a largely steeper “catch-up” decline since about 1960. The neglected concept in actuarial and epidemiological authors highlight that the near linear decline in research. Seligman et al., (2015) is one of the few to mortality rates for women at age 65 years fails to set out a detailed description of positive health as exhibit signs of “the introduction of antibiotics, the “well-being beyond the mere absence of disease” start of the National Health Service, or the changing and as the study of empirically measured health prevalence of tobacco smoking”. Instead, they assets and their configuration in an individual. In ascribe the decline to “a cumulative process driven this sense, a health asset refers to a factor that by rising living standards” and thereby consequent “produces longer life, lower morbidity, lower health improvements in positive health. Improvement in care expenditure, better prognosis when illness living standards, however, may act as a proxy or does strike, and/or higher quality of life” (Seligman coincide with changes in a range of public health et al., 2015). This concept is similar to that for related exposures, including improvements in mental health whereby it is no longer considered childhood nutrition and environmental exposures the same as the absence of mental illness, but such as air quality. Therefore we should not expect consists of a “measurable configuration of positive too much from aggregated statistics for mortality emotions, engagement, good relationships, rates in later life to reveal signals that unravel or meaning, and accomplishment” (Seligman 2011). distinguish various mechanisms that may be In contrast, life course epidemiology is a well- operating across the life course. established approach and defined as “the study of The integration of both social and biological long term effects on later health or disease risk of pathways provides new insights into aetiology of physical or social exposures during gestation, health and disease but brings about new challenges childhood, adolescence, young adulthood, and later in our views on generalisability of findings across

188

COMMENT AND DEBATE historical time, geographical regions or social strata compromised due to survival bias, but also any of a society. Although findings demonstrating a information on past exposures and experiences, large degree of consistency in early life either objective or subjective, can only, by determinants of social trajectories and later health definition be obtained retrospectively, with all the have recently emerged, for example, from research well-known limitations of that method. Current on several generations of Swedish men and women longitudinal and life course data analysis (Goodman, Gisselmann & Koupil, 2010), the provide a means to handle the interpretation of these effects within societies that problem of the survival bias (e.g., inverse dramatically change over time is not probability of attrition weights) (Weuve et al., 2012) straightforward. Would consistency in social but no such remedies can be developed to pathways indicate a failure of our social and public overcome the problems of retrospective data health policies to break the chains of social and collection. biological risks that affect subsequent generations? It remains, however, that healthy octogenarians Or would it perhaps indicate a change in the main and even more so healthy or even underlying pathophysiological (biological) ‘’, can be seen as living proof of mechanisms? successful ageing (Motta et al., 2005; Terry et al., In spite of the dramatic changes experienced 2008; Vacante et al., 2012; Willcox et al., 2008), and across the life course by the golden generation, it is positive health may prove to be a valuable their uniqueness in terms of biological and social framework to understand the explanatory factors at history that arguably limits their explanatory role work. Attempts have been made to understand here. The timespan covered by the decline in which life course trajectories led to the observed mortality rates at 65 years applies to people born exceptional longevity, for instance, using the many decades both prior and after that generation. ‘compression of morbidity’ hypothesis, that posits Each of these other cohorts, especially younger that in order to achieve extreme old age individuals men and women born in the decades since 1940, markedly delay or even escape diseases that would would need to have an equivalent biological and otherwise be lethal at younger ages (Evert, Lawler, social history with additional benefits on positive Bogan & Perls, 2003). Based on retrospective health in each decade. Further, longevity risk lifetime data on diagnoses of 10 major illnesses appears to have been an international (i.e., , heart disease, diabetes, stroke, phenomenon, so explanations are needed both non-skin cancer, osteoporosis, thyroid condition, over time and across populations. Yet without data Parkinson’s disease, and chronic obstructive from longitudinal studies, such accounts – including disease) and one ocular disease (cataracts) those from the context of UK social history provided collected by means of health questionnaires by Blane and colleagues on the golden generation – completed by >400 centenarians (or their proxies), would still be prone to ecological fallacies because the authors reported that three morbidity profiles they do not rely on empirical data of biological and fitted their study population: survivors, delayers and social variables obtained at the individual level escapers (Evert et al., 2003). Interestingly, the throughout the years. identification of these distinct phenotypes led them Since its inception, the Newcastle 85+ study used to conclude that the exceptional longevity of by Blane and colleagues to illustrate the golden centenarians may be achieved by multiple generation, has provided valuable information pathways and that there may be sex differences in about their physical, metabolic, immune and these. cognitive function phenotypes (Collerton et al., Blane and colleagues acknowledge the need to 2007; Harrison et al., 2015; Motta et al., 2005; look for evidence with other direct indicators of Terry, Sebastiani, Andersen & Perls, 2008; Vacante positive health. Seligman et al., (2015) specifically et al., 2012; Willcox et al., 2008). Yet by their suggest a re-analysis of existing longitudinal design, this and numerous other studies conducted datasets shifting from the traditional focus on ‘what worldwide among the oldest old are limited in their goes wrong in life’ to the study of health assets, contribution to the understanding of the life course including subjective ones, as the main instruments trajectories that explain the longevity of their to operationalize Positive Health. Some examples of participants. Not only their validity may be biological markers of positive health as proposed by

189

COMMENT AND DEBATE

Seligman et al., (2015) are low blood pressure (BP), monitoring of health at younger ages (Maggi, Irwin, low body mass index (BMI) and high Siddiqi & Hertzman, 2010) and ideally, one could cardiorespiratory fitness (VO2max). Some on-going use these as indicators of positive health early in life longitudinal studies have examined these variables to construct more continuous life course throughout the life course of individuals. For trajectories of positive health, rather than rely on instance, using a life course approach of repeated measurements taken at a later stages of the life data obtained between the ages of 12 and 36 years course which would be already influenced by a rate (8 repeated measures), the Amsterdam Growth and of decline. Health Longitudinal Study (AGAHLS) has shown that Aspects of self-report health such as SF-36 36-year olds with a healthy arterial ageing physical functioning scores, may provide another phenotype (i.e., younger than that that could be measure of positive health (Seligman et al., 2015; expected on the basis of their chronological age) Ware, Snow, Kosinski & Gandek, 1993). With (Ferreira & van de Laar, 2015) and metabolic progressive improvements in positive health, one profiles (Ferreira, Twisk, van Mechelen, Kemper & might expect higher scores at the same age with Stehouwer, 2005) were characterised by less steep each subsequent generation, and each with its own increases in BP, BMI and VO2max from adolescence distinct trajectory. The Australian Longitudinal to adulthood, with levels of BP and BMI never Study on Women’s Health has studied over 45,000 exceeding those cut-off values conventionally women in three age cohorts (born in 1921-26, placing individuals at higher risk for metabolic and 1946-51, and 1973-78) and has been running since cardiovascular diseases (Ferreira, van de Laar, Prins, 1996 (Dobson et al., 2015). Although the age ranges Twisk & Stehouwer, 2012). In addition, the AGHALS do not yet overlap to provide a definitive answer, has pin-pointed adolescence and transition from each cohort does show a higher set of scores for adolescence to young adulthood as critical periods physical functioning for younger generations (figure when increases in the levels of BP and BMI or 1), but the overall impression can still be decreases in VO2max may no longer be reversible in interpreted as a consistent connected trajectory some individuals. across life (and across these generations of The development of life course epidemiology women), rather than three distinct and and the increasing interest in potential critical or progressively higher trajectories. With further data sensitive periods have also stimulated a wider use in coming years the character of this life course of developmental and growth indicators for trajectory will be more clearly resolved.

190

Page 24 of 25

1 2 3 4 COMMENT AND DEBATE 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Review Only 19 20 21 22 23 24 25 26 27 28 29 30 31 Figure 1. SF-36 physical functioning scores (means with 95% confidence limits) by age and cohort 32 SF-36 physicalfor all women who provided survey data in the Australian Longitudinal Study on Women’s Health functioning scores (means with 95% confidence limits) by age and cohort for all women who 33 provided survey data at the specified age. 34 at the specified age. 169x127mm (96 x 96 DPI) 35 36 In summary, Blane and colleagues present heterogeneity across cohorts as (i) evidence against 37 intriguing evidence to support the case for an universal biological effects or (ii) evidence for 38 additional role of positive health as an explanation contextual effects and effect modification. 39 for longevity risk. To identify the specific Nevertheless, on-going longitudinal studies are 40 mechanisms operating across life and evident in the already applying the life course approach to the 41 42 biological and social history of each cohort, they better understanding of disease, health, and 43 acknowledge the next step needed is to move positive health, and hold the promise that it will not 44 beyond aggregated statistics for disease prevalence be too long until more definitive evidence to 45 and mortality, to evidence from direct measures of explain longevity risk can be identified – sufficient 46 health assets and across multiple cohorts and follow-up time just needs to accrue. It may still be a 47 populations. Yet for more detailed investigation in decade or more, but what are decades when trying 48 future research, interpretation of results from to understand the factors operating over 100-year 49 50 cross-cohort comparisons can be problematic as it timescales that led to this black swan event. 51 may not be obvious whether one would interpret 52 53 54 55 References 56 Collerton, J., Barrass, K., Bond, J., Eccles, M., Jagger, C., James, O., Martin-Ruiz, C., Robinson, L., von Zglinicki, 57 T., & Kirkwood, T. (2007). The Newcastle 85+ study: biological, clinical and psychosocial factors 58 associated with healthy ageing: study protocol. BMC Geriatrics, 7(14). 59 http://dx.doi.org/10.1186/1471-2318-7-14 60 Dobson, A., Hockey, R., Brown, W., Byles, J., Loxton, D., McLaughlin, D., Tooth, L., & Mishra, G.D. (2015). Cohort profile update: Australian Longitudinal Study on Women’s Health. International Journal of Epidemiology 44 (5),1547-1547f. http://dx.doi.org/10.1093/ije/dyv110

191

COMMENT AND DEBATE

Evert, J., Lawler, E., Bogan, H., & Perls, T. (2003). Morbidity profiles of centenarians: survivors, delayers, and escapers. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 58(3), 232-237. 58(3), 232-237. http://dx.doi.org/10.1093/gerona/58.3.M232 Ferreira, I., Twisk, J.W., van Mechelen, W., Kemper, H.C., & Stehouwer, C.D. (2005). Development of fatness, fitness, and lifestyle from adolescence to the age of 36 years: determinants of the metabolic syndrome in young adults: the Amsterdam Growth and Health Longitudinal Study. Archives of Internal Medicine, 165(1), 42-48. http://dx.doi.org/10.1001/archinte.165.1.42 Ferreira, I., & van de Laar, R.J. (2015). Lessons from the Amsterdam Growth and Health Longitudinal Study. In P.M. Nilsson, S. Laurent & M. H. Olsen (Eds.), Early Vascular Ageing – new directions in cardiovascular protection (pp. 33-44). Amsterdam: Elsevier Academic Press. http://dx.doi.org/10.1016/b978-0-12-801387-8.00005-3 Ferreira, I., van de Laar, R.J., Prins, M. H., Twisk, J.W., & Stehouwer, C.D. (2012). Carotid stiffness in young adults: a life-course analysis of its early determinants: the Amsterdam Growth and Health Longitudinal Study. Hypertension, 59(1), 54-61. http://dx.doi.org/10.1161/HYPERTENSIONAHA.110.156109 Goodman, A., Gisselmann, M.D., & Koupil, I. (2010). Birth characteristics and early-life social characteristics predict unequal educational outcomes across the life course and across generations. Data from a Swedish cohort born 1915-1929 and their grandchildren born 1973-1980. Longitudinal and Life Course Studies, 1(4), 317-338. Harrison, S.L., Stephan, B.C., Siervo, M., Granic, A., Davies, K., Wesnes, K.A., Kirkwood, T.B., Robinson, L., & Jagger, C. (2015). Is there an association between metabolic syndrome and cognitive function in very old adults? The Newcastle 85+ Study. Journal of the American Geriatric Society, 63(4), 667-675. http://dx.doi.org/10.1111/jgs.13358 Kuh, D., Ben-Shlomo, Y., Lynch, J., Hallqvist, J., & Power, C. (2003). Life course epidemiology. Journal of Epidemiology and Community Health, 57(10), 778-83. http://dx.doi.org/10.1136/jech.57.10.778 Maggi, S., Irwin, L.J., Siddiqi, A., & Hertzman, C. (2010). The social determinants of early child development: an overview. Journal of Paediatrics and Child Health, 46(11), 627-35. http://dx.doi.org/10.1111/j.1440-1754.2010.01817.x Mishra, G., Nitsch, D., Black, S., De Stavola, B., Kuh, D., & Hardy, R. (2009). A structured approach to modelling the effects of binary exposure variables over the life course. International Journal of Epidemiology, 38(2), 528-37. http://dx.doi.org/10.1093/ije/dyn229 Mishra, G.D., Chiesa, F., Goodman, A., De Stavola, B., & Koupil, I. (2013). Socio-economic position over the life course and all-cause, and circulatory diseases mortality at age 50-87 years: results from a Swedish birth cohort. European Journal of Epidemiology, 28(2), 139-47. http://dx.doi.org/10.1007/s10654-013-9777-z Motta, M., Bennati, E., Ferlito, L., Malaguarnera, M., Motta, L., & Italian Multicenter Study on, Centenarians. (2005%0. Successful aging in centenarians: myths and reality. Archives of Gerontology and Geriatrics, 40(3), 241-251. Seligman, M.E., Peterson, C., Barsky, A.J., Boehm, J.K., Kubzansky, L.D., Park, N., & Labarthe, D. (2015). Positive Health and Health Assets: Re-analysis of Longitudinal Datasets. Whitepaper. Retrieved 1 October 2015, 2015, from http://positivehealthresearch.org/content/positive-health-and-health- assets-re-analysis-longitudinal-datasets Seligman, M.E.P. (2011). Flourish. New York: Free Press. Taleb, N,N. (2007), The black swan: the impact of the highly improbable. New York: Random House. Terry, D.F., Sebastiani, P., Andersen, S. L., & Perls, T.T. (2008). Disentangling the roles of disability and morbidity in survival to exceptional old age. Archives of Internal Medicine, 168(3), 277-283. http://dx.doi.org/10.1001/archinternmed.2007.75 Vacante, M., D'Agata, V., Motta, M., Malaguarnera, G., Biondi, A., Basile, F., Malaguarnera, M., Gagliano, C., Drago, F., & Salamone, S. (2012). Centenarians and supercentenarians: a black swan. Emerging social, medical and surgical problems. BMC Surgery, 12 Suppl 1, S36. http://dx.doi.org/10.1186/1471-2482-12-S1-S36

192

COMMENT AND DEBATE

Ware, J. E., Snow, K.K., Kosinski, M., & Gandek, B. (1993). SF- 36 Health survey manual and interpretation guide. Boston: The Health Institute, New England Medical Center. Weuve, J., Tchetgen Tchetgen, E.J,, Glymour, M.M., Beck, T.L., Aggarwal, N.T., Wilson, R.S., Evans, D.A., & Mendes de Leon, C.F. (2012). Accounting for bias due to selective attrition: the example of smoking and cognitive decline. Epidemiology, 23(1), 119-128. http://dx.doi.org/10.1097/EDE.0b013e318230e861 Willcox, D.C., Willcox, B.J., Wang, N-C., He, Q., Rosenbaum, M., & Suzuki, M. (2008). Life at the extreme limit: phenotypic characteristics of supercentenarians in Okinawa. Journal of Gerontology: Medical Sciences, 63(11), 1201-1208. http://dx.doi.org/10.1093/gerona/63.11.1201 Willets, R., Gallop, A., Leandro, P., Lu, J., Macdonald, A., Miller, K., Richards, S., Robjohns, N., Ryan, J., Waters, H. (2004). Longevity in the 21st century. London: Institute of Actuaries & Faculty of Actuaries.

193

COMMENT AND DEBATE

Response by David Blane ESRC International Centre for Life Course Studies in Society Health and University College London, UK [email protected] Bola Akinwale ESRC International Centre for Life Course Studies in Society and Health and Imperial College London, UK Rebecca Landy ESRC International Centre for Life Course Studies in Society and Health and Queen Mary College London, UK Katey Matthews ESRC International Centre for Life Course Studies in Society and Health and University of Manchester, UK Morten Wahrendorf ESRC International Centre for Life Course Studies in Society and Health and University of Duesseldorf, Germany

The authors thank their six colleagues for When reading the commentaries on our piece, thoughtful comments. These come from we realised that we should say a few more words psychology (Hans-Werner Wahl), sociology (Mark about the Strachan-Sheikh model and our idea of Hayward), demography (Aart Liefbroer) and positive health. epidemiology (Gita Mishra, Isabel Ferreira, Ilona Koupil). It is heartening to see the level of The Strachan-Sheikh model agreement between them about required future We use the word model in the sense of a developments in life course research, particularly theoretical description of the way a system or the need for inter-disciplinarity, international process works; or, if preferred, a way of thinking comparative research and new studies to collect about reality. We accept readily the comments that combined high quality social and biological the different components of the anatomical- measurements; as pioneered, we add, by Richard physiological-biochemical system, which comprise Suzman’s legacy of the USA National Institute of the human body, grow and develop and lose Aging’s international portfolio of studies of ageing capacity and atrophy at varying rates. (in relation to the inclusion of biological Nevertheless, we find the generalised Strachan- measurements, particularly the English Longitudinal Sheikh model useful for two reasons. First, it Study of Ageing and, it is intended, the Survey of reminds us that humans are both social and Health, Ageing and Retirement in Europe). biological beings who, biologically, start as a single While celebrating this consensus, we regret the cell which sub-divides and replicates many times, to lack of comment from biology and actuarial science; each person’s own achieved peak capacity which and note that none of our commentators engaged subsequently attenuates at varying speeds to fully with either the biological or the social death. Second, it reminds us that these rates of historical parts of the ideas we presented. For us, it growth and atrophy are influenced by social is important that data collection and analysis should circumstances, material and emotional, producing be guided explicitly by social and biological variation in life expectancy and mean height and plausibility; so, for example, while agreeing that “… age at puberty which vary by country, social class important factors related to longevity risk seem not and historical period. It is this sensitivity of biology to belong to the health sphere”, we nevertheless to social context which makes social history consider that death is inescapably a biological event important to life course research, because it can and that social influences can produce death only inform us about who was likely to have been through biological processes. Similarly, we feel that exposed to what during which phase of their knowledge of social-historical context development, with some exposures being beneficial complements, rather than replaces, formal analysis and others noxious and the biological impact of of age-period-cohort effects. In other words, a specific exposures varying, to some extent, commitment to social and biological plausibility can depending on whether the person was in the stage help to move us along the road from statistical of growth or decline. association (…related to...) towards causation and the social challenge of longevity risk.

194

COMMENT AND DEBATE

In relation to the decline phase of the Strachan- diseases of childhood (scarlet fever, measles, Sheikh model, it is worth mentioning a recent study diphtheria, whooping cough, diarrhoea & vomiting) of healthy, athletic people aged 55-79 years who fell at the same time by proportionately similar spent two days completing a battery of amounts. Medical innovation (diphtheria anti- physiological measurements, of which only six toxin) and the public health sanitary reforms (water measures showed an inverse association with age & fly-borne infections) contributed to this change, (Pollock et al., 2015). These six involved mainly but so did increased host resistance primarily due tissue repair (ILGF-1, pelvic bone density) and lung to better nutrition (the difference in host resistance function (FEV1, VO2max, VT), which could be seen as contributes to high child mortality due to measles in components of intrinsic ageing or as signs of pre- refugees camps compared with lower mortality clinical morbidity. In either case, the main from the same virus in adequately nourished implication of this small scale, cross-sectional study children). Our idea of positive health is that it is the is that the biological path of decline to death is equivalent of host resistance for chronic, non- mainly via disease, rather than intrinsic ageing. If communicable diseases. replicated, these findings argue for greater Positive health in the sense of the body’s ability attention to variation in the rates of decline due to to neutralise or mitigate the effect of the living and working conditions of adult life. environmental insults may also underlie the secular increase in the population’s mean height and the Positive health fall in its mean age at the onset of puberty. And, we suggest in our piece, that the more or less linear The financial consequences of the recent and, fall across the twentieth century in the mortality particularly among men, unanticipated increase in rate of 65 year old women (see our figure 1) may be life expectancy at middle age, and its knock-on a further manifestation, due to cumulative effects at later ages, have led to efforts to improvements in the conditions of life, to which understand its causes. Actuaries are responsible for public health undoubtedly contributed, alongside predicting future changes in life expectancy, which technical-scientific innovation, democratic they are doing primarily in terms of anticipated competition for power and self-organisation in change in disease risk factors and therapeutic trade unions, pressure groups and so forth. This innovations, to which our piece suggested the suggestion raises two interesting questions in addition of a third factor, namely, changes to relation to historical change in mortality at age 65 positive health. years. Why did mortality rates at age 65 remain We were led to the idea of positive health by two largely stable for both men and women from the features of the present phenomenon and by the 1840s to 1900? And why did the fall in men’s previous occurrence of something similar. For us, it mortality rate at 65 plateau during the 1940s, ‘50s is unlikely to be a coincidence that the mortality and ‘60s? Amartya Sen’s aphorism that “There’s rates of the main causes of men’s death have never been a famine in a democracy” (Sen, 1994) declined by proportionately similar amounts during may be relevant to the former question, because the same period of time (see table 1 of our piece). the cohort who were aged 65 years in 1900, when Medical innovation, in our view, is unlikely to be the mortality among 65 year olds started to fall, were cause of this phenomenon, because its born in 1835, which was one year after the start of effectiveness varies greatly by disease – moderate the reform of the UK parliamentary electoral to high in the case of ischaemic heart disease; low franchise, and so were the first cohort of 65 year for carcinoma of the stomach. Similarly for risk olds to have lived their lives in an expanding factor change, where the most effective change democracy. To suggest an answer to the second (tobacco smoking cessation) impacts variably on question, we need to ask how the lives of men who different causes of death: high for lung cancer; low were aged 65 years during the 1940s-60s differed for accidents & violence. from women of the same age. Both were born The difficulty of explaining the present during the latter part of the nineteenth century, so phenomenon solely in terms of medical innovation the men were the right age to have fought in the and risk factor change reminded us of the late first world war; perhaps their later mortality nineteenth and early twentieth century when the mortality rates of the most prevalent infectious

195

COMMENT AND DEBATE plateau was an unrecognised long-term effect of develop, although we give perhaps greater their experiences during that war. emphasis than they do to the inclusion of biologists In summary, we welcome the enthusiasm of our in our inter-disciplinary efforts and the value of commentators for what we regard as the right social-biological plausibility as a means of moving direction in which life course research should from correlation to causation.

Acknowledgements MW is supported by funding from the German Research Foundation (Deutsche Forschungemeinschaft), project number WA 3065/3-1;DB is supported by UK Economic and Social Research Council award number ES/J019119/1

References Pollock, R., Carter, S., Vellose, C., Duggal, N., Lord, J., Lazarus, N., & Harridge, S. (2015). An investigation into the relationship between age and physiological function in highly active older adults. Journal of Physiology, 593, 657-680. Sen, A. (1994). Freedoms and needs. The New Republic, 10th January 1994, pp. 31-37.

196